第一部分 FastDFS介绍
1.FastDFS是什么
FastDFS是一款类Google FS的开源分布式文件系统,它用纯C语言实现,支持Linux、FreeBSD、AIX等UNIX系统。它只能通过 专有API对文件进行存取访问,不支持POSIX接口方式,不能mount使用。
准确地讲,Google FS以及FastDFS、mogileFS、 HDFS、TFS等类Google FS都不是系统级的分布式文件系统,而是应用级的分布式文件存储服务。
FastDFS的作者是余庆,现在淘宝网Java中间件团队从事Java基础平台研发工作。
以上文字引用自<<程序员>> 文章 <<分布式文件系统FastDFS架构剖析>>http://www.programmer.com.cn/tag/fastdfs-架构/
2.FastDFS的体系结构
3.FastDFS工作过程
1. Client询问Tracker server上传到的Storage server;
2. Tracker server返回一台可用的Storage server,返回的数据为该Storage server的IP地址和端口;
3. Client直接和该Storage server建立连接,进行文件上传,Storage server返回新生成的文件ID,文件上传结束。
1. Client询问Tracker server可以下载指定文件的Storage server,参数为文件ID(包含组名和文件名);
2. Tracker server返回一台可用的Storage server;
3. Client直接和该Storage server建立连接,完成文件下载。
FastDFS从V1.20开始,支持通过HTTP协议下载文件。
1.用户浏览器访问Tracker server内置的Web Server,URL中包含文件ID(包含组名和文件名);
2.Tracker server将这个HTTP请求redirect到一台可用的Storage server的Web Server上(可以是apache或nginx);
3.用户浏览器直接与Storage server的Web Serverr建立连接,完成文件下载。
第二部分 FastDFS+nginx_module配置手记
1.系统环境
最小化安装的 CentOS 5.6 x86_64
tracker server: 192.168.3.220
storage server: 192.168.3.226
2.环境准备[所有服务器上操作]
为了速度,配置使用sohu的源
cd /etc/yum.repo.d/
wget http://mirrors.sohu.com/help/CentOS-Base-sohu.repo
安装EPEL源
rpm -ihv http://download.fedora.redhat.com/pub/epel/5/x86_64/epel-release-5-4.noarch.rpm
安装开发工具和编译nginx所需的开发包
yum groupinstall "Development Tools"
yum install libevent-devel.x86_64 pcre-devel.x86_64 zlib-devel.x86_64
3.在tracker上的操作[安装tracker server]
[[email protected]tracker]# wget http://fastdfs.googlecode.com/files/FastDFS_v3.05.tar.gz
[[email protected]tracker]# tar xvzf FastDFS_v3.05.tar.gz
[[email protected]tracker]# cd FastDFS
[[email protected]tracker]# vi make.sh
将如下两行前边的"#"删除
WITH_HTTPD=1
WITH_LINUX_SERVICE=1
[[email protected]tracker]# ./make.sh
[[email protected]tracker]# ./make.sh install
[[email protected]tracker]# vi /etc/fdfs/tracker.conf
修改/etc/fdfs/tracker.conf,主要修改以下两处,如有其他调整,可参考文档自行调整,默认的配置也可以工作。
# the base path to store data and log files
base_path=/home/tracker ====> 放置data和log的目录
##include http.conf ====> #include http.conf 这里一定要注意!是 #include,不是include!!!
[[email protected]tracker]# mkdir -p /home/tracker
启动tracker服务
[[email protected]tracker]# /etc/init.d/fdfs_trackerd start
确认8080,22122端口已经监听
[[email protected]tracker]# netstat -nl|grep -E ‘8080|22122‘
tcp 0 0 0.0.0.0:22122 0.0.0.0:* LISTEN
tcp 0 0 0.0.0.0:8080 0.0.0.0:* LISTEN
4.在storage上的操作
[[email protected]storage]# wget http://fastdfs.googlecode.com/files/FastDFS_v3.05.tar.gz
[[email protected]storage]# wget http://fastdfs.googlecode.com/files/fastdfs-nginx-module_v1.10.tar.gz
[[email protected]storage]# wget http://nginx.org/download/nginx-1.0.11.tar.gz
[[email protected]storage]# tar xvzf FastDFS_v3.05.tar.gz
[[email protected]storage]# cd FastDFS
[[email protected]storage]# ./make.sh
[[email protected]storage]# ./make.sh install
[[email protected]storage]# tar xvzf nginx-1.0.11.tar.gz
[[email protected]storage]# tar xvzf fastdfs-nginx-module_v1.10.tar.gz
[[email protected]storage]# cd nginx-1.0.11
[[email protected]storage]# ./configure --prefix=/usr/local/nginx --add-module=/root/fastdfs-nginx-module/src
[[email protected]storage]# make
[[email protected]storage]# make install
[[email protected]storage]# vi /etc/fdfs/storage.conf
修改/etc/fdfs/storage.conf,主要修改以下几处,如有其他调整,可参考文档自行调整,默认的配置也可以工作。
# the name of the group this storage server belongs to
group_name=group1 ====> 此台storage server所属的服务器组名
# the base path to store data and log files
base_path=/home/storage ====> 放置data和log的目录
# store_path#, based 0, if store_path0 not exists, it‘s value is base_path
# the paths must be exist
store_path0=/home/storage ====> 放置文件的目录
# tracker_server can ocur more than once, and tracker_server format is
# "host:port", host can be hostname or ip address
tracker_server=192.168.3.220:22122 ====> tracker server的ip和端口,此处可以写多个tracker server,每行一个
#HTTP settings
http.disabled=true ====> 关闭内置的web server
# the port of the web server on this storage server
http.server_port=80 ====> web server的端口改成80
[[email protected]storage]# cp /root/fastdfs-nginx-module/mod_fastdfs.conf /etc/fdfs/
[[email protected]storage]# vi /etc/fdfs/mod_fastdfs.conf
# the base path to store log files
base_path=/home/storage ====> 放置log的目录
# FastDFS tracker_server can ocur more than once, and tracker_server format is
# "host:port", host can be hostname or ip address
tracker_server=192.168.3.220:22122 ====> tracker server的ip和端口,此处可以写多个tracker server,每行一个
# the group name of storage server
group_name=group1 ====> 此台storage server所属的服务器组名
# if uri including group name
# default value is false
url_have_group_name = true ====> 在URL中包含group名称
# store_path#, based 0, if store_path0 not exists, it‘s value is base_path
# the paths must be exist
store_path0=/home/storage ====> 放置文件的目录
[[email protected]storage]# mkdir -p /home/storage
在nginx的server配置段中增加M00的location声明
[[email protected]storage]# vi /usr/local/nginx/conf/nginx.conf
location /group1/M00 {
root /home/storage/data;
ngx_fastdfs_module;
}
创建M00目录的软连接
[[email protected]storage]# ln -s /home/storage/data /home/storage/data/M00
启动storage服务
[[email protected]storage]# /etc/init.d/fdfs_storaged start
启动nginx
[[email protected]tracker]# /usr/local/nginx/sbin/nginx
确认8080,22122端口已经监听
[[email protected]storage]# netstat -nl|grep -E ‘80|23000‘
tcp 0 0 0.0.0.0:23000 0.0.0.0:* LISTEN
tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN
5.在tracker上的操作[作为client测试]
修改/etc/fdfs/client.conf文件,主要修改以下几处,如有其他调整,可参考文档自行调整,默认的配置也可以工作。
# the base path to store log files
base_path=/tmp
# tracker_server can ocur more than once, and tracker_server format is
# "host:port", host can be hostname or ip address
tracker_server=192.168.3.220:22122
创建一个用于测试的文件demofile.txt
[[email protected]tracker]# vi demofile.txt
内容就一行字: dddddddddd
使用自带的fdfs_test上传文件测试。
[[email protected]tracker]# /usr/local/bin/fdfs_test /etc/fdfs/client.conf upload demofile.txt
This is FastDFS client test program v3.05
Copyright (C) 2008, Happy Fish / YuQing
FastDFS may be copied only under the terms of the GNU General
Public License V3, which may be found in the FastDFS source kit.
Please visit the FastDFS Home Page http://www.csource.org/
for more detail.
[2012-01-20 14:36:04] INFO - base_path=/tmp, connect_timeout=30, network_timeout=60, tracker_server_count=1, anti_steal_token=0, anti_steal_secret_key length=0
tracker_query_storage_store_list_without_group:
server 1. group_name=group1, ip_addr=192.168.3.226, port=23000
group_name=group1, ip_addr=192.168.3.226, port=23000
storage_upload_by_filename
group_name=group1, remote_filename=M00/00/00/wKgD4k8ZC1SNK-NYAAAAC3lx3Rk484.txt
source ip address: 192.168.3.226
file timestamp=2012-01-20 14:36:04
file size=11
file crc32=2037505305
file url: http://192.168.3.220:8080/group1/M00/00/00/wKgD4k8ZC1SNK-NYAAAAC3lx3Rk484.txt
storage_upload_slave_by_filename
group_name=group1, remote_filename=M00/00/00/wKgD4k8ZC1SNK-NYAAAAC3lx3Rk484_big.txt
source ip address: 192.168.3.226
file timestamp=2012-01-20 14:36:04
file size=11
file crc32=2037505305
file url: http://192.168.3.220:8080/group1/M00/00/00/wKgD4k8ZC1SNK-NYAAAAC3lx3Rk484_big.txt
6.使用浏览器打开上传的文件
我们在浏览器中输入http://192.168.3.220:8080/group1/M00/00/00/wKgD4k8ZC1SNK-NYAAAAC3lx3Rk484.txt
可以看到,访问的地址已经被redirect到了storage server的IP(192.168.3.226)了。